Given a named function, how can I call this function, and not the subfunction of the same name?

2015 5 29
3 个回答
回答已采纳
更新时间：2015 5 30
9 次查看（30 天）

0 个投票

My function is passed the name of a function (as a string). It just so happens that my function contains a subfunction of that name. I want to make sure I call the external function, instead of the subfunction. How can I achieve this?
E.g. save the following in test.m:
function test
feval(evalin('base', 'str2func(''help'')'), 'help');
end
function varargout = help(varargin)
error('Should not get here!');
[varargout{1:nargout}] = deal([]);
end
Then calling test gives:
>> test
Error using test>help (line 6)
Should not get here!
Error in test (line 2)
feval(evalin('base', 'str2func(''help'')'), 'help');

1 个评论
显示 -1更早的评论 隐藏 -1更早的评论

Oliver Woodford
Oliver Woodford 2015-5-29
I would say what I'm seeing is unexpected behaviour. Also, putting
evalin('base', 'which(''help'')')
as the first line in the help subfunction gives:
C:\....\test.m (help) % Subfunction of test
which is unexpected too.

请先登录，再进行评论。

请先登录，再回答此问题。

 采纳的回答

Titus Edelhofer
Titus Edelhofer 2015-5-29

1 个投票

Oliver,
what about this:
feval(evalin('base', '@(x) help(x)'), 'help')
This works fine. And as you noted yourself, using {:} you can expand the input variable as cell array.
Titus

更多回答（2 个）

Philip Borghesani
Philip Borghesani 2015-5-29
编辑：Philip Borghesani 2015-5-29

2 个投票

There is a much simpler solution to this:
function fh=test
fh=str2func('@(x) help(x)');
fh('help')
end
...
str2func isolates any anonymous function created from the surrounding environment see Loren's Blog

1 个评论
显示 -1更早的评论 隐藏 -1更早的评论

Oliver Woodford
Oliver Woodford 2015-5-30
Nice! Thanks, Phil. I find it bizarre that str2func finds a different function depending on whether you use the anonymous function syntax or just the function name, though.

请先登录，再进行评论。

Titus Edelhofer
Titus Edelhofer 2015-5-29

0 个投票

Hi Oliver,
although this is not an answer to your question it might help anyway: this is one of the reasons to use function handles instead of strings denoting functions. The big advantage of a function handle is, that the function is determined in the moment the function handle in contrast to strings, where in the moment of evaluation the dispatching happens.
If you pass @help instead of 'help' to your function as input, you are sure, that the correct function is used.
Titus

5 个评论
显示 3更早的评论 隐藏 3更早的评论

Oliver Woodford
Oliver Woodford 2015-5-29
Thanks, Titus. I know about function handles, but in my case they won't work. My function is compiling a mex file. The mex file has an associated m file which contains the information required to compile the mex. So depending on whether the mex file has been compiled, calling the function by name will call either the m file or the mex file. This is the desired effect. Using a handle would not allow this to work.
Titus Edelhofer
Titus Edelhofer 2015-5-29
Hm, so renaming the subfunctions to something highly unlikely to be of the same name as outside function does not work either :(.
Titus Edelhofer
Titus Edelhofer 2015-5-29
Oliver,
what about this:
feval(evalin('base', '@(x) help(x)'), 'help')
This works fine. The "only" drawback: you need to know the number of input and output variables to build up the string ...
Titus
Oliver Woodford
Oliver Woodford 2015-5-29
编辑：Oliver Woodford 2015-5-29
The unlikely function name thing is a workaround which I can use. I'm hoping I won't need to, though. I can use the solution above :). Although the number of input arguments aren't known, I can put them in a cell array:
feval(evalin('base', '@(x) help(x{:})'), {'help'})
If you make it an answer I'll accept it.
Alfonso Nieto-Castanon
编辑：Alfonso Nieto-Castanon 2015-5-29
couldn't you use:
feval(evalin('base','@(varargin) help(varargin{:})'),'help')
to account for variable number of inputs?

请先登录，再进行评论。

类别

在 帮助中心 和 File Exchange 中查找有关 Startup and Shutdown 的更多信息

产品

标签

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by